The standard vocabulary defines the valid statement syntax for Muss and parses syntax into special Rust Iterators which can be used to execute the statement.
To declare your own vocabulary, use Interpretor::with or Interpreter::with_vocab with a custom vocabulary (I'm not sure why you would, but I'm not going to stop you).
; -- The REPL will automatically place semicolons when Enter is pressed and it's not inside of brackets or a literal. Muss requires semicolons at the end of every statement, though, so Muss files must use semicolons.
Operations to reduce the items in an iterable: `iterable.(filter);`.
Filters are statements of the format `something.(predicate)`, where "something" is a variable name or another valid statement, and "predicate" is a valid filter predicate (see below).
E.g. `files(folder="~/Music/", recursive=true).(title == "Romantic Traffic");` is valid filter syntax to filter all songs in the Music folder for songs named "Romantic Traffic" (probably just one song).
Compare all items, keeping only those that match the condition. Valid field names change depending on what information is available when the Item is populated, but usually title, artist, album, genre, track, filename are valid fields. Optionally, a ? or ! can be added to the end of the field name to skip items whose field is missing/incomparable, or keep all items whose field is missing/incomparable (respectively).
Keep only the items that are at the start index up to the end index. Start and/or end may be omitted to start/stop at the iterable's existing start/end (respectively). This stops once the end condition is met, leaving the rest of the iterator unconsumed.
Keep only the items that are at the start index up to and including the end index. Start may be omitted to start at the iterable's existing start. This stops once the end condition is met, leaving the rest of the iterator unconsumed.
#### predicate1 || predicate2 -- e.g. `iterable.(4 || 5);`
Keep only the items that meet the criteria of predicate1 or predicate2. This will always consume the full iterator.
#### [empty] -- e.g. `iterable.();`
Matches all items
#### if filter: operation1 else operation2 -- e.g. `iterable.(if title == "Romantic Traffic": repeat(item, 2) else item.());`
Replace items matching the filter with operation1 and replace items not matching the filter with operation2. The `else operation2` part may be omitted to preserve items not matching the filter. To perform operations with the current item, use the special variable `item`. The replacement filter may not contain || -- instead, use multiple filters chained together.
Keep only items which are do not duplicate another item, or keep only items whoes specified field does not duplicate another item's same field. The first non-duplicated instance of an item is always the one that is kept.
Functions are statements of the format `function_name(params)`, where "function_name" is one of the function names (below) and params is a valid parameter input for the function.
Each function is responsible for parsing it's own parameters when the statement is parsed, so this is very flexible.
E.g. `files(folder="~/Music/", recursive=true);` is valid function syntax to execute the files function with parameters `folder="~/Music/", recursive=true`.
Initialize the SQLite database connection using the provided parameters. This must be performed before any other database operation (otherwise the database will already be connected with default settings). This returns an empty iterable (contains zero items).
Combine multiple iterables in an interleaved pattern. This is a variant of union(...) where the first item in iterable1, then iterable2, ... is returned, then the second item, etc. until all iterables are depleted. There is no limit to the amount of iterables which can be provided as parameters.
Combine multiple iterables in a sequential pattern. All items in iterable1 are returned, then all items in iterable2, ... until all provided iterables are depleted. There is no limit to the amount of iterables which can be provided as parameters.
Combine multiple iterables such that only items that exist in iterable1 and iterable2 and ... are returned. The order of items from iterable1 is maintained. There is no limit to the amount of iterables which can be provided as parameters.
Sort by an Item field. Valid field names change depending on what information is available when the Item is populated, but usually title, artist, album, genre, track, filename are valid fields. Items with a missing/incomparable fields will be sorted to the end.
Shuffle the songs in the iterator. This is random for up to 2^16 items, and then the randomness degrades (but at that point you won't notice). The more verbose syntax is allowed in preparation for future randomisation strategies.
Sort by musical similarity, starting with a random first song from the iterator. The optional qualifier may be `chroma`, `loudness`, `spectrum`, or `tempo`. When the qualifier is omitted, they are all considered for comparing audio similarity.
Sort by the distance (similarity) from the first song in the iterator. Songs which are more similar (lower distance) to the first song in the iterator will be placed closer to the first song, while less similar songs will be sorted to the end. This uses the [bliss music analyser](https://github.com/polochon-street/bliss-rs), which is a very slow operation and can cause music playback interruptions for large iterators. This requires the `advanced` feature to be enabled (without the feature enabled this is still valid syntax but doesn't change the order).
Sort by the distance (similarity) between the last played song in the iterator. Similar to bliss_first. The song which is the most similar (lower distance) to the previous song in the iterator will be placed next to it, then the process is repeated. This uses the [bliss music analyser](https://github.com/polochon-street/bliss-rs), which is a very slow operation and can cause music playback interruptions for large iterators. This requires the `advanced` feature to be enabled (without the feature enabled this is still valid syntax but doesn't change the order).
Comma-separated procedure steps will be executed sequentially (like a for loop in regular programming languages). The variable item contains the current item of the iterable.
Declare the variable and (optionally) set the initial value to something. The assignment will only be performed when the variable has not yet been declared. When the initial value (and equals sign) is omitted, the variable is initialized as empty().
Branch based on a boolean condition. Multiple comma-separated procedure steps may be supplied in the if and else branches. This does not currently support if else chains, but they can be nested to accomplish similar behaviour.
An iterable operation inside of the procedure. When assigned to item, this can be used to replace item with multiple others. Note that iterable operations are never executed inside the procedure; when item is iterable, it will be executed immediately after the end of the procedure for the current item.
Format a value into a string. This behaves differently depending on the value's type: When the value is an Item, the item's corresponding field will replace all `{field}` instances in the format string. When the value is a primitive type (String, Int, Bool, etc.), the value's text equivalent will replace all `{}` instances in the format string. When the value is an iterable operation (Op), the operation's script equivalent will replace all `{}` instances in the format string.
